Help required for a novel recording technology

[SallyFord]

Hi all,

I am currently studying for my MSc in Forensic Archaeological Science. For my dissertation I am investigating a mock cremation crime scene using polynomial texture mapping, and comparing it to traditional recording. The analysis will be done using a survey undertaken by individuals from a range of backgrounds, to simulate expert witnesses, a jury and students or junior crime scene investigators. Although this project is specifically related to forensics, it has a wider potential for archaeology (currently being explored at The University of Southampton).
I would be very grateful if some of you could take some time to have a look at my website before carrying out the survey. All feedback and comments are welcome, and everybody is invited to take part.

https://sites.google.com/site/sallyfordu...t-proposal

Thanks

Sally

[Unitof1]

This technology is still limited however. It cannot measure the features within the image, so traditional measurements would still be required in order to produce an accurate record. Neither does it produce a geographical location, and so traditional methods would also be required for this

I dont understand why this technique cannot do these things, surely if you know your camera position and orientation and your light sources you should be able to measure and locate "features" The fact that the techinque produces images does mean that features are "interpretations" but by taking a "feature" away and rephotographing that object should have its place in your record.

PS I like the use of the word map

[Jack]

Yep, rectified photography is what your looking for.

Also if you are trying to interpret a 'site' without excavation your missing often vital evidence. Like subtle context differences, such as different layers, or cut features and important relationships between them. These can only be gathered through careful dismantling of the site. Like layers with differing compaction often only discernible by feel would indicated prolonged use/reuse or several episodes of deposition.

You will also potentially miss those little details that point to the evidence/solution you are looking for that are often discovered during excavation.

Think of the case of a fire in a house. The house owner claims criminals broke in and set a fire in the kitchen. On the face of it, recorded by photos, the glass in the kitchen door is broken, the door open. The burn patterns indicate where the fire was started and that an accelerant was used. Case closed?

No. On closer (destructive) examination. Soot was found under all of the pieces of glass from the door, showing that the fire was started before the door glass was broken. The house owner was examined......traces of accelerant was found on their clothes and hands......

The same is the case for archaeology. A photo no matter the detail and lighting/ digital effects will not record as much information as a skilled excavation.

BUT
As another tool in the inventory the technique looks ok (if of limited use at the mo) The actual shape, location and measurements of a context, be it a pit, or a layer of charcoal is of paramount importance.
Maybe for digitally recording cleaned and conserved objects to help look for tool-marks, micro wear etc? But again measurements are important.

I would have thought that measurements would be important in forensics too?!??

[barkingdigger]

Hi Sally,

You need to look at metric photography techniques such as rectified photograhs and photogrammetry. Adding a set of surveyed control points in the images lets you measure within the final product. Kubit's PhoToPlan is one type of rectification software (runs in AutoCAD, so you can then trace off the image to "draw" the features as scale drawings), while Topcon has offered stereo photogrammetry software in the few-grand price range. Using your varied lights is interesting, but with survey points and a second set of pics to make stereo pairs you can add all this metric capability as & when you need it. After all, snapping extra pics is cheap!

PS: I'm not pushing either specific package mentioned above, just giving examples. There are other similar solutions out there commercially available! The point is you need to talk to archaeological surveyors and buildings folk because they use these and know the pros and cons of the different techniques.

[kevin wooldridge]

Speaking as an archaeological surveyor with a an interest in digital visualisation.....It strikes me that what you really need to be able to do is 1) Georeference your photos 2) Create an accurate point cloud of the surface and associated features 3) drape photos over the wireframe created from the point cloud. All of this is possible using current technology. Laser scanner with built in video facility, camera with inbuilt GPS, good visualisation software...Not sure about the PTM technology, but if it produces images it would work with the technology outlined above..although I don't see it coming in cheap....which also seemed to be one of your considerations. And it would still involve some form of physical involvement to record the relationships between the various interfaces recorded....An interesting idea for a project though.

[Unitof1]

https://www.hpl.hp.com/research/ptm/

getting a feeling that this method is thrown at smallish objects to show up relief, not sure if any rectification involves a lot of fisheye and that using average digital cameras would result in resolution/file size problems for bigger areas particularly ones without an obviouse centre of attention

[Bodger51]

Metric photography is interesting, but if your looking at it for applications and stable datasets, the integration of light source with Laser scanners should be a good focus.

But if your looking at high resolution detail photography, thats a lot of setup for laser imagery, whilst enmass applications would be dogmatic, unless particular details are becoming evident for courtcase applications, other than within robust cross-examination considerations by second or additional assessments.

[Bodger51]

Survey data to imagery cross referencing has other elements at play, but thats at tech. level detail

[Bodger51]

The idea holds good as experiental crime scene re-enactments, but your going deeper than public discussions could facilitate.